Firefly Video Model: Adobe’s new AI tool to generate videos from text

Adobe has launched its Firefly Video Model, an AI tool that generates video from text prompts, stepping into the growing competition in generative AI for film and television production. This move positions Adobe alongside rivals like OpenAI, ByteDance, and Meta Platforms, all of whom have recently released similar video tools.

Adobe, however, distinguishes itself by training its models on data it owns the rights to, ensuring the generated content can be legally used for commercial purposes. While a general release date is not confirmed, Adobe has begun offering access to those who signed up for the waiting list.

Although no customers have been announced for the video tool, Gatorade is using Adobe’s image generation model to create custom bottle designs, and Mattel has applied the technology in designing packaging for its Barbie dolls. Adobe has geared its video tools towards creators, making them user-friendly for everyday video production.

Ely Greenfield, Adobe’s chief technology officer for digital media, highlighted that the focus is on ensuring the AI understands key video production concepts like camera angles and motion, allowing it to seamlessly blend with conventional footage.